JP4301091B2 - Acoustic signal encoding device - Google Patents

Acoustic signal encoding device Download PDF

Info

Publication number
JP4301091B2
JP4301091B2 JP2004184829A JP2004184829A JP4301091B2 JP 4301091 B2 JP4301091 B2 JP 4301091B2 JP 2004184829 A JP2004184829 A JP 2004184829A JP 2004184829 A JP2004184829 A JP 2004184829A JP 4301091 B2 JP4301091 B2 JP 4301091B2
Authority
JP
Japan
Prior art keywords
information
frequency band
band
acoustic signal
amount
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
JP2004184829A
Other languages
Japanese (ja)
Other versions
JP2006010817A (en
Inventor
孝朗 山辺
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Victor Company of Japan Ltd
Original Assignee
Victor Company of Japan Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Victor Company of Japan Ltd filed Critical Victor Company of Japan Ltd
Priority to JP2004184829A priority Critical patent/JP4301091B2/en
Publication of JP2006010817A publication Critical patent/JP2006010817A/en
Application granted granted Critical
Publication of JP4301091B2 publication Critical patent/JP4301091B2/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Description

本発明は音響信号符号化装置に係り、特に情報圧縮によって生じる量子化雑音を的確に制御し、音質を保ちながら高速に符号化信号を生成する音響信号符号化装置に関する。   The present invention relates to an acoustic signal encoding apparatus, and more particularly to an acoustic signal encoding apparatus that accurately controls quantization noise generated by information compression and generates an encoded signal at high speed while maintaining sound quality.

時間領域のディジタル化された音響信号を周波数領域に変換した後、周波数領域にて情報圧縮を行って符号化信号を生成する音響信号符号化装置では、ビデオＣＤ等で用いられるMPEG-1 AUDIO Layer2、インターネット等のデータ配信で利用されているMPEG-1 Layer3（略称ＭＰ３）、ＢＳディジタル放送で採用されているMPEG-2 AUDIO NBC、別名アドバンスト・オーディオ・コーディング（Advanced Audio Coding：略称ＡＡＣ）、ＤＶＤ（Digital Versatile Disk）の音声フォーマットであるドルビーディジタル（Dolby Digital）、ＭＤ（ミニディスク）の圧縮方式ＡＴＲＡＣまたはＡＴＲＡＣ３などの符号化方式で符号化を行う。   In an audio signal encoding apparatus that generates an encoded signal by compressing information in the frequency domain after converting the digitized audio signal in the time domain to the frequency domain, MPEG-1 AUDIO Layer 2 used in a video CD or the like. MPEG-1 Layer3 (abbreviated as MP3) used in data distribution over the Internet, MPEG-2 AUDIO NBC used in BS digital broadcasting, also known as Advanced Audio Coding (abbreviated as AAC), DVD Encoding is performed by an encoding method such as Dolby Digital or MD (minidisk) compression method ATRAC or ATRAC3, which is an audio format of (Digital Versatile Disk).

これらは時間領域のディジタル化された音響信号を周波数領域に変換し、特定の周波数帯に偏りを持つ音響信号の特徴と、人間の聴覚の特性を考慮した聴覚の感度に応じた周波数帯毎の重み付けに従い、聴感的に重要とされない周波数帯の情報を削減又は減少することによって情報圧縮を行っている。   They convert time domain digitized acoustic signals to frequency domain, and each frequency band according to the sensitivity of the acoustic signal considering the characteristics of the acoustic signal biased in a specific frequency band and the characteristics of human hearing. According to the weighting, information compression is performed by reducing or reducing information in frequency bands that are not audibly important.

上記の符号化方式において、特にＭＰ３やＡＡＣは周波数帯毎の情報量削減方法に二重の繰り返し処理を取り入れており、非常に細かい尺度で周波数帯毎の情報量を制御している。上記の二重の繰り返し処理のうちの一方の繰り返し処理は、圧縮信号の品質を向上するための個々の周波数帯毎における量子化雑音を制御するものであり、もう一方の繰り返し処理は圧縮率に従ったブロック毎の割り当て情報量以内で符号化を行うために符号化情報量を制御するものである。両者の条件を満足するために、一方の繰り返し処理が終了した後、他方の繰り返し処理へと移行し、相互に渡り合う事によって品質及び符号化情報量の要求を両立している。   In the above encoding method, in particular, MP3 and AAC incorporate double repetition processing into the information amount reduction method for each frequency band, and control the information amount for each frequency band on a very fine scale. One of the above-described double iterative processes controls the quantization noise for each frequency band to improve the quality of the compressed signal, and the other iterative process determines the compression rate. The encoded information amount is controlled to perform encoding within the allocated information amount for each block. In order to satisfy both conditions, after one repetitive process is completed, the process proceeds to the other repetitive process, and both requirements of quality and encoded information amount are satisfied.

ところが、これらの繰り返し処理は非常に負荷が重く、ディジタル信号処理プロセッサ（ＤＳＰ）や中央処理装置（ＣＰＵ）などを用いたリアルタイム性が重要なシステムに対してふさわしいとはいえない。従って、リアルタイム処理を必要とする用途に対しては、制限時間で符号化を終了するための負荷の軽い量子化部が求められる。   However, these repetitive processes are very heavy and cannot be said to be suitable for systems in which real-time performance using a digital signal processor (DSP) or a central processing unit (CPU) is important. Therefore, for applications that require real-time processing, a light-weight quantizing unit for ending encoding within a limited time is required.

そこで、本来の二重の繰り返し処理の代わりに符号化情報量を制御する一重の繰り返し処理のみを行い、量子化雑音の制御は最初の１回の処理で終了する構造を取り入れ高速化を実現する音響信号符号化装置が従来より知られている（例えば、特許文献１参照）。この従来の音響信号符号化装置の動作について、図６のフローチャートと共に説明するに、まず、周波数帯重み付け情報によって各々の周波数帯域の量子化精度を修正するための個別帯域量子化精度を算出する（ステップＳ１）。   Therefore, instead of the original double iterative process, only a single iterative process that controls the amount of encoded information is performed, and the quantization noise control is realized by adopting a structure that ends in the first one process. An acoustic signal encoding device is conventionally known (for example, see Patent Document 1). The operation of this conventional acoustic signal encoding apparatus will be described with reference to the flowchart of FIG. 6. First, the individual band quantization accuracy for correcting the quantization accuracy of each frequency band is calculated based on the frequency band weighting information ( Step S1).

次に、この個別帯域量子化精度と全帯域共通量子化精度の初期値を用いて、入力音響信号である周波数信号を量子化器において所定の演算式に基づき量子化を行う（ステップＳ２）。続いて、量子化後の周波数信号を集計し、他の符号化に必要な補助情報と共に総情報量を算出する（ステップＳ３）。続いて、求められた総情報量が情報圧縮率に依存するブロックに対して与えられた割り当て情報量以内であるかどうか判定する（ステップＳ４）。   Next, using the initial values of the individual band quantization accuracy and the all-band common quantization accuracy, the frequency signal that is the input acoustic signal is quantized in a quantizer based on a predetermined arithmetic expression (step S2). Subsequently, the quantized frequency signals are totaled, and the total amount of information is calculated together with auxiliary information necessary for other encoding (step S3). Subsequently, it is determined whether or not the obtained total information amount is within the allocation information amount given to the block depending on the information compression rate (step S4).

量子化後の総情報量が与えられた割当情報量以内と判定された場合、符号化品質及び符号化情報量の両者の条件を満たすものとして、量子化を終了する。他方、量子化後の総情報量が与えられた割り当て情報量を超えていると判定された場合、全帯域共通量子化精度を一律に粗くする方向に修正し、量子化雑音が増加することによる若干の符号化品質劣化を許容しながら符号化情報量の条件を満たすため、新しい全帯域共通量子化精度値を設定し直す（ステップＳ５）。   When it is determined that the total amount of information after quantization is within the given allocation information amount, the quantization is terminated assuming that both the encoding quality and the encoded information amount are satisfied. On the other hand, if it is determined that the total amount of information after quantization exceeds the assigned information amount, all-band common quantization accuracy is corrected so as to be uniformly rough, and quantization noise increases. In order to satisfy the condition of the encoded information amount while allowing a slight deterioration in encoding quality, a new all-band common quantization accuracy value is reset (step S5).

次に、この設定し直した全帯域共通量子化精度と個別帯域量子化精度とを用いて、入力音響信号である周波数信号を量子化器において所定の演算式に基づき再度量子化を行う（ステップＳ２）。以下、ステップＳ３の量子化後の周波数信号の集計を行った後、符号化によって生成された総情報量が割り当て情報量の条件を満足するまで全帯域共通量子化精度値を変更する処理を繰り返す。このようにして、全体の量子化精度を加減しながら目標とする符号化情報量へ調整するための繰り返し処理を経て符号化信号を生成する。   Next, using the re-set all-band common quantization accuracy and individual-band quantization accuracy, the frequency signal that is the input acoustic signal is quantized again in the quantizer based on a predetermined arithmetic expression (step) S2). Thereafter, after the frequency signals after quantization in step S3 are aggregated, the process of changing the all-band common quantization accuracy value is repeated until the total information amount generated by the encoding satisfies the condition of the allocated information amount. . In this way, an encoded signal is generated through an iterative process for adjusting to the target encoded information amount while adjusting the overall quantization accuracy.

特開２００１−１４８６３２号公報JP 2001-148632 A

しかしながら、上述の一重の繰り返し処理による量子化部を有する従来の音響信号符号化装置は、二重の繰り返し処理による量子化部を有する音響信号符号化装置に比べて、符号化情報量を制御する一重の繰り返し構造を採用しているため処理量を削減でき、高速に符号化信号を生成することはできるが、全帯域共通量子化精度情報の初期値が適当でない場合、繰り返し処理の回数が増大し、必ずしも高速であるとはいえない場合がある。   However, the conventional acoustic signal encoding device having the quantization unit based on the single iteration process described above controls the amount of encoded information as compared with the acoustic signal encoding device having the quantization unit based on the double iteration process. Since the single repetition structure is adopted, the amount of processing can be reduced and the encoded signal can be generated at high speed. However, if the initial value of all-band common quantization accuracy information is not appropriate, the number of repetition processing increases. However, it may not always be fast.

また、全帯域共通量子化精度情報を初期値から変更する事によって、各々の帯域において最適であった個別の量子化精度に誤差が生じる可能性があり、符号化情報量は目標とすべき情報量を満足するものの、理想的な量子化雑音レベルからの乖離が起きてしまうことで音質が低下するといった危惧がある。   In addition, by changing the all-band common quantization accuracy information from the initial value, an error may occur in the individual quantization accuracy that is optimal in each band, and the amount of encoded information should be the target information Although the amount is satisfied, there is a concern that the sound quality deteriorates due to the deviation from the ideal quantization noise level.

本発明は以上の点に鑑みなされたもので、本来量子化部において二重の繰り返し構造を持つ音響信号符号化方式を高速処理が可能な構造にするべく、全帯域共通量子化精度に予め適当と推定される数値を与え、それに基づき量子化雑音が許容される水準へと個別帯域量子化精度値を修正し、割り当て情報量については打ち切り周波数帯域を設ける事によって符号化情報量の制御を行うことで、負荷の軽い音響信号符号化装置を提供することを目的とする。   The present invention has been made in view of the above points, and is suitable in advance for all-band common quantization accuracy in order to make an acoustic signal coding system originally having a double repetition structure in a quantization unit capable of high-speed processing. Based on this, the individual band quantization accuracy value is corrected to a level where quantization noise is allowed, and the amount of allocated information is controlled by providing a truncation frequency band. Thus, an object of the present invention is to provide an acoustic signal encoding device with a light load.

本発明は上記の目的を達成するため、時間領域のディジタル化された音響信号をある定められた時間間隔でブロック化した後、ブロック内の音響信号を複数の周波数帯域に分割し、分割した各周波数帯域毎の重み付けに基づき、ブロック内の音響信号の量子化を行い情報圧縮する音響信号符号化装置において、複数の周波数帯域からなる全周波数帯域共通の全帯域共通量子化精度値を決定する全帯域共通量子化精度決定部と、分割した各周波数帯域毎の聴覚的な重み付けによって、全帯域共通量子化精度値を各周波数帯域毎に修正するための個別帯域量子化精度値を決定する個別帯域量子化精度決定部と、全帯域共通量子化精度値及び個別帯域量子化精度値に従い、ブロック内の音響信号を量子化する量子化器と、情報圧縮率及びディジタル化された音響信号の特性に従い、ブロック毎に情報量を割り当てる割り当て情報量決定部と、量子化器から出力された量子化信号の符号化情報量を集計し、集計した符号化情報量に基づいて量子化信号の周波数帯域を決定する符号化情報量集計・周波数帯域決定部とを備え、符号化情報量集計・周波数帯域決定部は、量子化器から出力された量子化信号における分割した各周波数帯域のうち、最も低域の分割周波数帯域の符号化情報量を集計し、情報量決定部によって割り当てられた情報量と最も低域の分割周波数帯域の符号化情報量とを比較し、最も低域の分割周波数帯域の符号化情報量が割り当てられた情報量より小さければ最も低域の分割周波数帯域に基づく周波数帯域を仮決定し、仮決定した周波数帯域が適当であるか否かをその周波数帯域の情報量が不足であるか否かにより判断し、適当でなければ次に高域の分割周波数帯域へ周波数帯域を広げて、その広げた周波数帯域における符号化情報量を集計し、その広げた周波数帯域における符号化情報量が割り当てられた情報量以内で、かつ、広げた周波数帯域が適当であると判断されるまで、最も低域の分割周波数帯域から高域の分割周波数帯域まで順次に周波数帯域を広げて、周波数帯域を最終決定することを特徴とする。 In order to achieve the above-mentioned object, the present invention blocks a digitized acoustic signal in a time domain at a predetermined time interval, and then divides the acoustic signal in the block into a plurality of frequency bands. In an acoustic signal encoding apparatus that quantizes an acoustic signal in a block and compresses information based on the weighting for each frequency band, the entire bandwidth common quantization accuracy value that is common to all frequency bands consisting of a plurality of frequency bands is determined. An individual band that determines an individual band quantization accuracy value for correcting the entire band common quantization accuracy value for each frequency band by an auditory weighting for each divided frequency band and a common band quantization accuracy determination unit a quantization accuracy determining unit, in accordance with the total bandwidth common quantization precision value and the individual band quantization precision values, a quantizer for quantizing the acoustic signal in the block, information compression ratio and digital According characteristics of the acoustic signal, aggregates the assignment information amount determining unit to assign the amount of information for each block, the encoded information amount of the quantized signal outputted from the quantizer, based on the aggregated coded information amount A coded information amount totaling / frequency band determining unit that determines the frequency band of the quantized signal, and the encoded information amount totaling / frequency band determining unit is configured to divide each quantized signal output from the quantizer. Of the frequency bands, the amount of encoded information in the lowest divided frequency band is aggregated, and the amount of information allocated by the information amount determining unit is compared with the amount of encoded information in the lowest divided frequency band. If the encoded information amount of the low frequency division frequency band is smaller than the allocated information amount, a frequency band based on the lowest frequency division frequency band is provisionally determined, and whether or not the provisionally determined frequency band is appropriate is determined. Zhou Judgment is made based on whether the amount of information in several bands is insufficient. If not appropriate, the frequency band is expanded to the next higher frequency band, and the amount of encoded information in the expanded frequency band is tabulated. Sequentially from the lowest divided frequency band to the highest divided frequency band until the encoded information amount in the expanded frequency band is within the allocated information amount and the expanded frequency band is determined to be appropriate The frequency band is widened and the frequency band is finally determined .

この発明では、従来繰り返し処理が行われていた量子化部において、音質条件を満たすための全帯域共通量子化精度値の統計的及び適応的な推定を行う全帯域共通量子化精度決定部と、符号化情報量を満たすための周波数帯域決定部とにより、繰り返し処理を削減することができる。また、音響信号の特性と全帯域共通量子化精度値との誤差による量子化信号の符号化情報量の増減を、符号化情報量集計・周波数帯域決定部によって吸収することができる。
ここで、上記の符号化情報量集計・周波数帯域決定部は、最も低域の周波数帯域及び広げた周波数帯域における符号化情報量が割り当てられた情報量以内でない場合に、他のブロックで余剰した余剰情報量を補助情報として割り当てられた情報量に追加することを特徴とする。 In the present invention, in the quantization unit that has been repeatedly performed conventionally, the all-band common quantization accuracy determination unit that performs statistical and adaptive estimation of the all-band common quantization accuracy value to satisfy the sound quality condition, The iterative process can be reduced by the frequency band determining unit for satisfying the encoded information amount. Further, the increase / decrease of the encoded information amount of the quantized signal due to the error between the characteristics of the acoustic signal and the all-band common quantization accuracy value can be absorbed by the encoded information amount aggregation / frequency band determining unit .
Here, the coded information amount aggregation / frequency band determining unit described above is surplus in other blocks when the coded information amount in the lowest frequency band and the expanded frequency band is not within the allocated information amount. The surplus information amount is added to the information amount allocated as auxiliary information.

また、本発明は上記の目的を達成するため、全帯域共通量子化精度決定部は、ブロック内の音響信号エネルギー、音響信号振幅値及び割り当て情報量のうちの少なくとも一の情報を要素として予め求めた全帯域共通量子化精度値の統計情報に基づく全帯域共通量子化精度値と、現在符号化中の現ブロックの音響信号よりも過去のブロックの音響信号の符号化の際に用いた全帯域共通量子化精度値とのいずれか一方又は両者の組み合わせ値を、全帯域共通量子化精度値として決定することを特徴とする。この発明では、全帯域共通量子化精度値を繰り返し処理を行わずに決定することができる。   Further, in order to achieve the above object, the present invention determines in advance the all-band common quantization accuracy determination unit using at least one of the acoustic signal energy, the acoustic signal amplitude value, and the allocated information amount in the block as an element. All-band common quantization accuracy value based on the statistical information of all-band common quantization accuracy value and all bands used for encoding the acoustic signal of the past block rather than the acoustic signal of the current block being encoded One of the common quantization accuracy values or a combination value of both is determined as the all-band common quantization accuracy value. In the present invention, it is possible to determine the all-band common quantization accuracy value without performing iterative processing.

また、上記の目的を達成するため、本発明は全帯域共通量子化精度決定部を、現ブロックの音響信号のブロック内の音響信号エネルギー、最大または代表的な振幅値、及びブロック内割り当て情報量の何れか一つ又は二つ以上を組み合わせた情報を、現ブロックの音響信号から抽出する音響信号特性抽出部と、ブロック内の音響信号のエネルギー、最大または代表的な音響信号の振幅値、及びブロック割り当て情報量の何れか一つ又は二つ以上を組み合わせて得られた全帯域共通量子化精度値の統計結果を示す統計的音響信号特性分類情報と、音響信号特性抽出部から出力された情報とを比較し、統計的音響信号特性分類情報の中から音響信号特性抽出部から出力された情報に最も近似する統計的音響信号特性分類情報を検索する特性比較部及び近似情報検索部と、検索結果から現ブロックの全帯域共通量子化精度値を特定する全帯域共通量子化精度特定部とを有する構成としたことを特徴とする。この発明では、音響信号の特性を考慮した全帯域共通量子化精度値を取得することができる。   Further, in order to achieve the above object, the present invention sets the all-band common quantization accuracy determination unit to the acoustic signal energy in the block of the acoustic signal of the current block, the maximum or representative amplitude value, and the allocated information amount in the block. An acoustic signal characteristic extraction unit that extracts information combining any one or two or more from the acoustic signal of the current block, the energy of the acoustic signal in the block, the amplitude value of the maximum or representative acoustic signal, and Statistical acoustic signal characteristic classification information indicating statistical results of all-band common quantization accuracy values obtained by combining any one or more of the block allocation information amounts, and information output from the acoustic signal characteristic extraction unit Compares statistical sound signal characteristic classification information and searches for statistical sound signal characteristic classification information that most closely matches the information output from the acoustic signal characteristic extraction part. And fine approximate information retrieval unit, characterized by being configured to have a full-band common quantization accuracy specifying unit for specifying a full band common quantization precision values of the current block from the search results. According to the present invention, it is possible to acquire the all-band common quantization accuracy value in consideration of the characteristics of the acoustic signal.

また、上記の目的を達成するため、本発明は、全帯域共通量子化精度決定部を、過去に決定した全帯域共通量子化精度値を用いて量子化された量子化信号の符号化情報量と、過去の１又は複数ブロックの量子化信号の周波数帯域の情報とを再帰的に用いて、全帯域共通量子化精度特定部で特定された全帯域共通量子化精度値を補正する補正部を含む構成としたことを特徴とする。この発明では、音響信号の特性を考慮し、過去の情報を再帰的に用いた補正部により、信頼性の高い全帯域共通量子化精度値を得ることができる。   Further, in order to achieve the above object, the present invention provides an all-band common quantization accuracy determination unit for encoding information amount of a quantized signal quantized using an all-band common quantization accuracy value determined in the past. And a correction unit that corrects the all-band common quantization accuracy value specified by the all-band common quantization accuracy specifying unit by recursively using the frequency band information of the past one or a plurality of blocks of quantized signals. It is characterized by including a configuration. According to the present invention, a highly reliable all-band common quantization accuracy value can be obtained by a correction unit that recursively uses past information in consideration of the characteristics of an acoustic signal.

本発明によれば、従来繰り返し処理が行われていた量子化部において、音質条件を満たすための全帯域共通量子化精度値の統計的及び適応的な推定を行う全帯域共通量子化精度決定部と、符号化情報量を満たすための周波数帯域決定部とにより、繰り返し処理を削減し、また、音響信号の特性と全帯域共通量子化精度値との誤差による量子化信号の符号化情報量の増減を、適応的な情報割り当てと周波数帯域決定部とによって吸収することで符号化情報量の制御を容易に可能としたため、低負荷な構成により音質劣化を抑えた符号化信号を高速に生成することができる。   According to the present invention, in the quantization unit that has been repeatedly performed in the past, the all-band common quantization accuracy determination unit that performs statistical and adaptive estimation of the all-band common quantization accuracy value for satisfying the sound quality condition And a frequency band determining unit for satisfying the encoded information amount, and reducing the iterative processing, and the encoded information amount of the quantized signal due to an error between the characteristics of the acoustic signal and the all-band common quantization accuracy value. Since the increase / decrease is absorbed by the adaptive information allocation and the frequency band determination unit, the amount of encoded information can be easily controlled, so that a low-load configuration generates a high-speed encoded signal with reduced sound quality degradation. be able to.

また、本発明によれば、音響信号の特性を考慮し、過去の情報を再帰的に用いた補正部によって信頼性の高い全帯域共通量子化精度値を特定するようにしたため、品質劣化をできるだけ抑えて音質を維持することができる。   In addition, according to the present invention, a highly reliable all-band common quantization accuracy value is specified by a correction unit that recursively uses past information in consideration of the characteristics of the acoustic signal, so that quality degradation can be minimized. Suppressing and maintaining sound quality.

次に、本発明を実施するための最良の形態について図面と共に説明する。図１は一般的な音響信号符号化装置のブロック図を示す。同図において、時間領域のディジタル音響信号は、時間−周波数変換部１に供給され、ある一定の時間間隔によってブロック化された後、ブロック内のディジタル音響信号が複数の周波数帯域に分割され周波数信号へと変換される。また、上記の時間領域のディジタル音響信号は、周波数帯重み付け情報算出部２に供給され、上記のブロック化された周波数信号の特徴を抽出するため、上記の分割された周波数帯域毎に人間の聴覚特性を用いた聴感的な感度を示す周波数帯重み付け情報が算出される。   Next, the best mode for carrying out the present invention will be described with reference to the drawings. FIG. 1 shows a block diagram of a general acoustic signal encoding apparatus. In the figure, a digital audio signal in the time domain is supplied to the time-frequency conversion unit 1 and is blocked at a certain time interval, and then the digital audio signal in the block is divided into a plurality of frequency bands. Converted to. In addition, the digital audio signal in the time domain is supplied to the frequency band weighting information calculation unit 2 to extract the characteristics of the blocked frequency signal, so that human auditory perception is performed for each of the divided frequency bands. Frequency band weighting information indicating auditory sensitivity using characteristics is calculated.

時間−周波数変換部１から出力された周波数帯域に変換された信号は、周波数帯重み付け情報算出部２から出力された周波数帯重み付け情報を基に、量子化部３にて周波数帯域毎に異なる量子化精度で量子化されて情報圧縮される。量子化部３にて得られた量子化信号は、符号化信号生成部４に供給され、ここで周波数帯重み付け情報や個別の周波数帯域量子化精度情報などからなる補助情報と多重化されて音響信号符号化信号が生成される。   Based on the frequency band weighting information output from the frequency band weighting information calculation unit 2, the signal converted into the frequency band output from the time-frequency conversion unit 1 is quantized by the quantization unit 3 for each frequency band. The information is compressed by quantization with the accuracy. The quantized signal obtained by the quantizing unit 3 is supplied to the encoded signal generating unit 4, where it is multiplexed with auxiliary information including frequency band weighting information, individual frequency band quantization accuracy information, etc. A signal encoded signal is generated.

ＡＡＣやＭＰ３などではこの量子化部３に前述のように繰り返し構造を採用している。そのため符号化品質は向上するが、符号化処理は非常に多くの演算が必要である。   In AAC, MP3, etc., the quantization unit 3 employs a repetitive structure as described above. Therefore, although the encoding quality is improved, the encoding process requires a large number of operations.

ＡＡＣの場合、量子化精度は全帯域共通量子化精度を表すグローバルゲイン（ｇｌｏｂａｌ＿ｇａｉｎ）と個別の周波数帯域量子化精度を表すスケールファクタ（Ｓｃａｌｅｆａｃｔｏｒ）を用いて次式で与えられている。   In the case of AAC, the quantization accuracy is given by the following equation using a global gain (global_gain) representing the all-band common quantization accuracy and a scale factor (Scalefactor) representing the individual frequency band quantization accuracy.

Ｘ＝(int)(sample3/4×２((Scalefactor-global_gain)×3/16))
ただし、上式中、Ｘは量子化後の周波数信号であり、Ｓａｍｐｌｅは量子化される周波数信号の絶対値であり、「（ｉｎｔ）」は浮動小数点で表現された数値から整数値への変換を示す。 X = (int) (sample 3/4 × 2 ((Scalefactor-global_gain) × 3/16) )
In the above equation, X is a frequency signal after quantization, Sample is an absolute value of the frequency signal to be quantized, and “(int)” is a conversion from a numerical value expressed in floating point to an integer value. Indicates.

上式中において、２のべき乗に関わるｇｌｏｂａｌ＿ｇａｉｎは（マイナス）の要素を持つがゆえ、大きければ大きいほど量子化される周波数サンプルの量子化後振幅を小さくする。この時、整数値に変換される量子化後の周波数信号Ｘの振幅が小さくなるに従い量子化精度は粗くなり、逆量子化後信号の量子化雑音が増大する事に繋がる。   In the above equation, global_gain related to the power of 2 has a (minus) element, so that the larger the value, the smaller the quantized amplitude of the frequency sample to be quantized. At this time, as the amplitude of the frequency signal X after quantization converted into an integer value becomes smaller, the quantization accuracy becomes coarser, leading to an increase in quantization noise of the signal after inverse quantization.

一方、一律の量子化精度を与えるｇｌｏｂａｌ＿ｇａｉｎに対し、Ｓｃａｌｅｆａｃｔｏｒは２のべき乗に対し（プラス）の要素を持ち、Ｓｃａｌｅｆａｃｔｏｒを増加する事によって、個別帯域の量子化後の周波数信号Ｘの振幅を増加させる方向に修正する事ができるので、個々の帯域の量子化精度を向上させるためにはＳｃａｌｅｆａｃｔｏｒの値を増やしてあげれば良い事がわかる。   On the other hand, with respect to global_gain that gives uniform quantization accuracy, the scale factor has a positive element with respect to the power of 2, and by increasing the scale factor, the amplitude of the frequency signal X after quantization of the individual band is increased. Since it can be corrected in the direction, it can be seen that the value of Scalefactor can be increased in order to improve the quantization accuracy of each band.

ここで、全帯域共通量子化精度値（ｇｌｏｂａｌ＿ｇａｉｎ）は、ブロック毎の符号化品質を左右する重要な要素である。符号化品質と符号化情報量との間では相反する関係がある。すなわち、符号化品質を高くするには符号化情報量が増加するが、符号化情報量には伝送路で定められた上限値があり、その許容範囲内で符号化品質をできるだけ高くするように、符号化品質と符号化情報量の両者の最適な関係を、前述した従来の二重の繰り返し処理により導くには多くの処理を費やさねばならない。   Here, the all-band common quantization accuracy value (global_gain) is an important factor that affects the coding quality of each block. There is a conflicting relationship between the encoding quality and the encoded information amount. In other words, the amount of encoded information increases to increase the encoding quality, but the encoded information amount has an upper limit determined by the transmission path, and the encoding quality should be as high as possible within the allowable range. In order to derive the optimum relationship between the encoding quality and the encoded information amount by the above-described conventional double repetitive processing, a lot of processing must be spent.

一方、リアルタイム処理が要求される状況においては、処理量の最大値がシステム全体のコストを押し上げる要因に直結し、できるだけ平均化された処理工程を持つ事が望まれる。従って、繰り返し構造は少なければ少ないほどリアルタイム処理にはふさわしい。   On the other hand, in a situation where real-time processing is required, it is desirable to have processing steps averaged as much as possible because the maximum value of the processing amount is directly linked to a factor that increases the cost of the entire system. Therefore, the fewer repeating structures, the more suitable for real-time processing.

また、全帯域共通量子化精度値（ｇｌｏｂａｌ＿ｇａｉｎ）はそもそもの音響信号の持つエネルギーや振幅値と依存関係を持ち、信号エネルギーが大きい、または音圧が高い時などは量子化精度を粗くしてもよく、更に割り当て情報量が少ない時も必然的に量子化精度を粗くし、情報量を節約しなくてはならいことは容易に推測できる。これは聴感特性からも裏付けられることであり、音響信号エネルギーや音圧が高い場合は、聴感的に許容される雑音レベルが増加し、よりマスキング効果が働くからである。   In addition, the quantization accuracy value (global_gain) common to all bands is dependent on the energy and amplitude value of the sound signal in the first place, and when the signal energy is large or the sound pressure is high, the quantization accuracy may be roughened. Well, it can be easily guessed that even when the amount of allocated information is small, the quantization accuracy is inevitably coarsened to save the information amount. This is also supported by the auditory characteristics. When the acoustic signal energy and the sound pressure are high, the noise level that is perceptually permissible increases and the masking effect is more effective.

音質を左右するのは情報圧縮によって各々の帯域に生じる量子化雑音の影響である。人間の聴感特性は低域ほど感度が良く、高域は非常に暖味である。よって、１０数ｋＨｚを超える帯域の瞬間的な存在・非存在の現象は捉える事が難しく、言い換えると周波数帯域が高域にて変化するような音質の違いを把握できないと推測される。   It is the influence of quantization noise generated in each band by information compression that affects the sound quality. The human auditory characteristics are more sensitive at lower frequencies and very warm at higher frequencies. Therefore, it is difficult to grasp the instantaneous presence / absence phenomenon in a band exceeding 10 kHz, in other words, it is assumed that the difference in sound quality that changes the frequency band in the high frequency range cannot be grasped.

また、ＡＡＣやＭＰ３は厳密には一定情報量が与えられる固定転送レートに従い符号化が行われるが、ブロック毎に情報量の貸し借りが許されており、擬似的な可変情報割り当てが可能である。   Strictly speaking, AAC and MP3 are encoded according to a fixed transfer rate to which a constant amount of information is given. However, lending and borrowing of the amount of information is allowed for each block, and pseudo variable information can be assigned.

以上の点に鑑み、本発明では音響信号エネルギー、音響信号振幅値及び割り当て情報量の何れか一つまたは組み合わせにより示される音響信号の特性によって全帯域共通量子化精度を推定し、更に符号化情報量は低域信号を優先した周波数帯符号割り当てによって制御する手段を講じたものである。   In view of the above points, in the present invention, all-band common quantization accuracy is estimated from the characteristics of the acoustic signal indicated by any one or a combination of acoustic signal energy, acoustic signal amplitude value, and allocated information amount, and further encoded information The amount is provided by means for controlling the frequency band code by giving priority to the low frequency signal.

図２は本発明になる音響信号符号化装置の一実施の形態における量子化部のブロック図を示す。図２において、図１の量子化部３に相当する量子化部１０は、量子化器１１、個別帯域量子化精度決定部１２、全帯域共通量子化精度決定部１３、割り当て情報決定部１４、符号化情報量集計部１５、周波数帯域決定部１６とからなる。全帯域共通量子化精度決定部１３による全帯域共通量子化精度の決定には数種類の方法が考えられる。ここでは音響信号の特性に応じて予め繰り返し処理を用いた量子化部で求めた全帯域共通量子化精度値を音響信号の特性別に分類したテーブルを使って導く手法を例に上げる。   FIG. 2 shows a block diagram of a quantization unit in an embodiment of an acoustic signal encoding apparatus according to the present invention. In FIG. 2, a quantization unit 10 corresponding to the quantization unit 3 in FIG. 1 includes a quantizer 11, an individual band quantization accuracy determination unit 12, an all-band common quantization accuracy determination unit 13, an allocation information determination unit 14, It comprises an encoded information amount totaling unit 15 and a frequency band determining unit 16. Several methods can be considered for the determination of the all-band common quantization accuracy by the all-band common quantization accuracy determination unit 13. Here, as an example, a method of deriving using a table in which all-band common quantization accuracy values obtained in advance by a quantizing unit using iterative processing according to the characteristics of the acoustic signal are classified according to the characteristics of the acoustic signal is used.

この本実施の形態の動作について、図３のフローチャートを併せ参照して説明する。図１に示した時間-周波数変換部１において生成された周波数信号が、図２の量子化器１１へ伝送される。同じく図１に示した周波数帯重み付け情報算出部２にて導かれた周波数帯重み付け情報が個別帯域量子化精度決定部１２へ伝送され、この情報を基に個別帯域量子化精度値が決定される（図３のステップＳ１１）。   The operation of this embodiment will be described with reference to the flowchart of FIG. The frequency signal generated in the time-frequency converter 1 shown in FIG. 1 is transmitted to the quantizer 11 in FIG. Similarly, the frequency band weighting information derived by the frequency band weighting information calculation unit 2 shown in FIG. 1 is transmitted to the individual band quantization accuracy determination unit 12, and the individual band quantization accuracy value is determined based on this information. (Step S11 in FIG. 3).

一方、全帯域共通量子化精度決定部１３では、入力される現ブロック音響信号の特性を基に、予め分類してある全帯域共通量子化精度値統計テーブルを用いて、現ブロックに近似する音響信号を検索し、その音響信号の特性に最もふさわしいと推定される統計上の全帯域共通量子化精度値を抽出する（図３のステップＳ１２）。なお、この全帯域共通量子化精度値は過去に用いられた量子化精度値を代入してもよい。この理由は音響信号が定常的な場合、量子化精度はブロック間で大きく異なる事がないという性質から来ている。   On the other hand, the all-band common quantization accuracy determination unit 13 uses the all-band common quantization accuracy value statistical table that has been classified in advance based on the characteristics of the input current block sound signal to approximate the current block. The signal is searched, and the statistical all-band common quantization accuracy value estimated to be most suitable for the characteristics of the acoustic signal is extracted (step S12 in FIG. 3). In addition, you may substitute the quantization precision value used in the past for this all-band common quantization precision value. This is because the quantization accuracy does not vary greatly between blocks when the acoustic signal is stationary.

更に、全帯域共通量子化精度に用いた推定値を、１又は複数の過去のブロック間の推移や、後述の割り当て帯域の増減、使用可能な情報配分量により調整する事によって、より信頼性の高い量子化精度値を取得する事が可能である。   Furthermore, by adjusting the estimated value used for all-band common quantization accuracy by the transition between one or more past blocks, the increase / decrease in allocated bandwidth, which will be described later, and the amount of usable information, more reliable It is possible to obtain a high quantization accuracy value.

量子化器１１は、上記の個別帯域量子化精度値と全帯域共通量子化精度値とを用いて、周波数信号を量子化信号へ変換する（図３のステップＳ１３）。この量子化信号は、例えば、前記式のＸで表わされ、前記式で示すように個別帯域量子化精度値（Ｓｃａｌｅｆａｃｔｏｒ）と、全帯域共通量子化精度値（ｇｌｏｂａｌ＿ｇａｉｎ）とが反映された信号である。変換された量子化信号は符号化情報量集計部１５へ供給される。   The quantizer 11 converts the frequency signal into a quantized signal using the individual band quantization accuracy value and the all-band common quantization accuracy value (step S13 in FIG. 3). This quantized signal is represented by, for example, X in the above equation, and a signal reflecting an individual band quantization accuracy value (Scalefactor) and an all-band common quantization accuracy value (global_gain) as shown in the above equation. It is. The converted quantized signal is supplied to the encoded information amount totaling unit 15.

符号化情報量集計部１５は、符号化総情報量を得るために、量子化信号と符号化に必要な補助情報から総情報量を算出する（図３のステップＳ１４）。総情報量は低域の周波数から順次算出し、ブロック毎に割り当てられた割り当て情報量を超えない帯域まで集計していく。   The encoded information amount totaling unit 15 calculates the total information amount from the quantized signal and auxiliary information necessary for encoding in order to obtain the encoded total information amount (step S14 in FIG. 3). The total amount of information is calculated sequentially from the low frequency range, and the total amount of information is added up to a bandwidth that does not exceed the allocated information amount assigned to each block.

一方、割り当て情報決定部１４は、符号化レート情報と周波数信号とを入力信号として受け、現ブロックで使用可能な割り当て情報量を算出する。この割り当て情報量は符号化レートによって自ずと求まるが、擬似的な可変レートが許されているＡＡＣやＭＰ３では、ブロック毎の音響信号の複雑度に応じて情報量をブロック間で増減する事によって、より高品質な符号化が実現できる。情報量の増減は入力信号のエネルギー、振幅値、変化量等によって決定される。   On the other hand, the allocation information determination unit 14 receives the coding rate information and the frequency signal as input signals, and calculates an allocation information amount usable in the current block. This allocation information amount is naturally determined by the coding rate, but in AAC and MP3 where a pseudo variable rate is allowed, by increasing or decreasing the information amount between blocks according to the complexity of the acoustic signal for each block, Higher quality encoding can be realized. The increase / decrease of the information amount is determined by the energy, amplitude value, change amount, etc. of the input signal.

以上の工程で得られた量子化信号は量子化雑音が分割周波数帯域毎に制御された状態であり、品質面では問題ないが、符号化情報量が割り当て情報量以内でなければならないという条件を満たす必要がある。そこで、周波数帯域決定部１６では、この符号化情報量の条件を満足するため、人間の聴感特性にとって重要な分割周波数帯域に優先的に情報を割り当てる周波数帯域の決定を行う（図３のステップＳ１５）。すなわち、周波数帯域決定部１６は、割り当て情報量以内である最高帯域を特定し、周波数帯域の制限によって符号化総情報量を制御する。   The quantized signal obtained in the above process is in a state where the quantization noise is controlled for each divided frequency band, and there is no problem in quality, but the condition that the encoded information amount must be within the allocated information amount. It is necessary to satisfy. Therefore, the frequency band determination unit 16 determines a frequency band that preferentially allocates information to the divided frequency band important for human auditory characteristics in order to satisfy the encoded information amount condition (step S15 in FIG. 3). ). That is, the frequency band determination unit 16 specifies the highest band that is within the allocated information amount, and controls the encoded total information amount by limiting the frequency band.

以上で、量子化部１０で行われる処理は終了し、最終的に得られた量子化信号とその補助情報が符号化信号生成部４において多重化され、符号化信号として生成される。   Thus, the processing performed by the quantizing unit 10 is completed, and the finally obtained quantized signal and its auxiliary information are multiplexed by the encoded signal generating unit 4 and generated as an encoded signal.

本実施の形態によれば、従来繰り返し処理が行われていた量子化部１０において、音質条件を満たすための量子化精度値の統計的及び適応的な推定を行う全帯域共通量子化精度決定部１３と、符号化情報量を満たすための周波数帯域決定部１６とにより、図３に示すように、繰り返し処理を削減し、非常に高速で、かつ、音質劣化を抑えた良好な音響信号符号化信号を生成する事ができる。   According to the present embodiment, in the quantization unit 10 that has conventionally been subjected to iterative processing, the all-band common quantization accuracy determination unit that performs statistical and adaptive estimation of the quantization accuracy value for satisfying the sound quality condition 13 and a frequency band determination unit 16 for satisfying the amount of encoded information, as shown in FIG. 3, it is possible to reduce the repetitive processing, achieve a very high speed, and suppress the deterioration of sound quality. A signal can be generated.

次に、図２中の個別のブロックについて更に詳しく説明する。全帯域共通量子化精度決定部１３による全帯域共通量子化精度決定方法の例として統計情報を用いる方法と、以前使用した量子化精度値を用いる方法と、使用した符号化情報量や周波数帯域の増減量などの補助情報を用いる方法について述べる。これらの方法は単独で使用しても組み合わせて使用してもよい。   Next, the individual blocks in FIG. 2 will be described in more detail. As an example of the all-band common quantization accuracy determination method by the all-band common quantization accuracy determination unit 13, a method using statistical information, a method using a previously used quantization accuracy value, and the amount of encoded information and frequency band used A method of using auxiliary information such as an increase / decrease amount will be described. These methods may be used alone or in combination.

まず、統計量から求める方法について説明する。既に簡単な説明をしたように、二重の繰り返し構造を持つ量子化部において量子化された音響信号は、品質及び情報量のバランスがとれた状態で量子化されており、仮に全帯域共通量子化精度値が既知なる値であれば、符号化情報量を制御するための繰り返し処理を省略できるはずである。   First, a method for obtaining from statistics will be described. As already explained, the acoustic signal quantized in the quantization section having a double repetition structure is quantized in a state where the quality and the amount of information are balanced. If the encoding accuracy value is a known value, iterative processing for controlling the amount of encoded information should be able to be omitted.

音響信号の特性から考慮して、この全帯域共通量子化精度値はブロック内音響信号エネルギー、音響信号振幅値、割り当て情報量に大きく関係している。そこで、予めこれらを要素とした全帯域共通量子化精度値の統計を取り、音響信号エネルギー、音響信号振幅値、割り当て情報量の何れか一つまたは組み合わせたとき、最も適当と推定される全帯域共通量子化精度が得られるようなテーブルを用意しておく事で、繰り返し処理によらずとも信頼性の高い全帯域共通量子化精度値を取得する事ができる。前記実施の形態の説明は、この方法による。   In consideration of the characteristics of the acoustic signal, the all-band common quantization accuracy value is largely related to the intra-block acoustic signal energy, the acoustic signal amplitude value, and the allocated information amount. Therefore, if the statistics of all-band common quantization accuracy values using these as elements are taken in advance and any one or a combination of acoustic signal energy, acoustic signal amplitude value, and allocated information amount is combined, all bands estimated to be most appropriate By preparing a table capable of obtaining the common quantization accuracy, it is possible to obtain a highly reliable all-band common quantization accuracy value without using the iterative process. The description of the embodiment is based on this method.

次に、以前使用した量子化精度値を用いる方法について説明する。この方法は、符号化過程で導いた過去のブロックにおける全帯域共通量子化精度値を現ブロックの全帯域共通量子化精度値として当てはめる方法である。音響信号が定常的に推移しているとき、情報圧縮で使用すべき量子化精度はブロック間で大きく異なる事はない。前ブロックで使用した全帯域共通量子化精度値を現ブロックにて使用しても、多くの場合、品質劣化には繋がらない。従って、音響信号が定常的であると判定された場合、前ブロックの全帯域共通量子化精度値をそのまま現ブロックの全帯域共通量子化精度値に用いる事によって、量子化処理を極めて簡単に行うことができる。   Next, a method using the previously used quantization accuracy value will be described. In this method, the all-band common quantization accuracy value in the past block derived in the encoding process is applied as the all-band common quantization accuracy value of the current block. When the acoustic signal is constantly changing, the quantization accuracy to be used for information compression does not vary greatly between blocks. Even if the all-band common quantization accuracy value used in the previous block is used in the current block, in many cases, it does not lead to quality degradation. Therefore, when it is determined that the acoustic signal is stationary, the quantization processing is extremely easily performed by using the entire block common quantization accuracy value of the previous block as it is as the entire block common quantization accuracy value. be able to.

次に、使用した符号化情報量や周波数帯域の増減量などの補助情報を用いる方法について説明する。この方法は、割り当て情報量又は符号化周波数帯域情報による全帯域共通量子化精度値推定方法であり、前記の二つの方法と組み合わせて用いるとより効果的である。統計的情報や過去のブロックで用いた全帯域共通量子化精度値を代用した場合、最終的に符号化した結果、予定より情報を多く使用したり、あるいは少なく使用したり、また、予定より周波数帯域が高かったり、あるいは低かったりと、推定による誤差が生じることがある。これは音響信号の変化によるものであり、現在符号化中の音響信号の特性に由来するものと考えられる。推定した全帯域共通量子化精度値と符号化中の音響信号との特性の不一致が原因であるため、これを修正する必要がある。修正する手段として、過去の割り当て情報量の大小や周波数帯域の高低などの情報をフィードバックすることによって推定誤差を小さくする事ができる。   Next, a method of using auxiliary information such as the amount of encoded information used and the amount of increase / decrease in frequency band will be described. This method is an all-band common quantization accuracy value estimation method based on the allocation information amount or the encoded frequency band information, and is more effective when used in combination with the above two methods. If the statistical information and the all-band common quantization accuracy value used in the past block are substituted, the final encoding results in using more or less information than planned, and using more frequency than planned. If the bandwidth is high or low, an estimation error may occur. This is due to a change in the acoustic signal, and is considered to be derived from the characteristics of the acoustic signal currently being encoded. This is due to the mismatch in characteristics between the estimated all-band common quantization accuracy value and the acoustic signal being encoded. This needs to be corrected. As a correction means, it is possible to reduce the estimation error by feeding back information such as the amount of past allocation information and the level of the frequency band.

図４は図２中の全帯域共通量子化精度決定部１３の一実施の形態のブロック図を示す。この全帯域共通量子化精度決定部１３は、これまで説明した全帯域共通量子化精度値算出方法を組み合わせて全帯域共通量子化精度を決定する回路部である。図４において、音響信号特性抽出部１３１は、入力された現ブロックの周波数信号のブロック内音響信号エネルギー、最大又は代表的な振幅値、及びブロック内割り当て符号量の何れか一つ又は任意の二つ若しくは三つを組み合わせた情報である音響信号の特性を算出（抽出）する。   FIG. 4 shows a block diagram of an embodiment of the all-band common quantization accuracy determination unit 13 in FIG. The all-band common quantization accuracy determination unit 13 is a circuit unit that determines the all-band common quantization accuracy by combining the all-band common quantization accuracy value calculation method described so far. In FIG. 4, the acoustic signal characteristic extracting unit 131 is any one of the in-block acoustic signal energy, the maximum or representative amplitude value, and the intra-block allocated code amount or any two of the input frequency signals of the current block. Calculate (extract) the characteristics of the acoustic signal, which is information combining one or three.

統計的音響信号特性分類情報１３２は、ブロック内音響信号エネルギー、最大又は代表的な振幅値、及びブロック内割り当て符号量の何れか一つ又は任意の二つ若しくは三つを組み合わせて得られた全帯域共通量子化精度値の統計結果を示す情報であり、予め外部の記憶装置等に記憶されている。   The statistical acoustic signal characteristic classification information 132 is obtained by combining any one or any two or three of the intra-block acoustic signal energy, the maximum or representative amplitude value, and the intra-block allocated code amount. This is information indicating the statistical result of the common band quantization accuracy value, and is stored in advance in an external storage device or the like.

特性比較器及び近似情報検索器１３３は、音響信号特性抽出部１３１からの現ブロックの音響信号の特性と統計的音響信号特性分類情報１３２とを比較し、統計的音響信号特性分類情報１３２の中で現ブロックの音響信号の特性に最も近似した統計的音響信号特性分類情報を検索する。この特性比較器及び近似情報検索器１３３により検索された統計的音響信号特性分類情報は、全帯域共通量子化精度特定器１３４に供給され、ここで、現ブロックにふさわしいと推定される全帯域共通量子化精度値が特定される。   The characteristic comparator / approximation information search unit 133 compares the acoustic signal characteristic of the current block from the acoustic signal characteristic extraction unit 131 with the statistical acoustic signal characteristic classification information 132, and stores the statistical acoustic signal characteristic classification information 132. To retrieve the statistical sound signal characteristic classification information that most closely approximates the sound signal characteristic of the current block. The statistical acoustic signal characteristic classification information searched by the characteristic comparator / approximation information searcher 133 is supplied to the all-band common quantization accuracy specifying unit 134, where all-band common estimated to be appropriate for the current block is used. A quantization accuracy value is identified.

全帯域共通量子化精度特定器１３４で特定された全帯域共通量子化精度値は、外部へ出力される一方、符号化情報蓄積器１３５に過去情報として蓄積される。符号化情報蓄積器１３５には、過去の周波数帯域情報や割り当て符号化情報量なども供給されて蓄積される。全帯域共通量子化精度特定器１３４は、符号化情報蓄積器１３５で蓄積された、以前のブロックにおいて使用された全帯域共通量子化精度値や周波数帯域情報、割り当て符号化情報量などが現在符号化中の音響信号との特性情報としてフィードバック入力され、特性比較器及び近似情報検索器１３３からの一般的な統計的音響信号特性分類情報を、フィードバック入力された現在符号化中の音響信号との特性との相違点に基づき修正し、最終的な全帯域共通量子化精度値を得る。   The all-band common quantization accuracy value specified by the all-band common quantization accuracy specifying unit 134 is output to the outside and is stored as past information in the encoded information storage unit 135. The encoded information storage unit 135 is supplied with and stores past frequency band information and the amount of allocated encoded information. The all-band common quantization accuracy specifying unit 134 stores the all-band common quantization accuracy value, the frequency band information, the allocated encoded information amount, etc. used in the previous block stored in the encoding information storage unit 135. Feedback as the characteristic information with the acoustic signal being encoded, and the general statistical acoustic signal characteristic classification information from the characteristic comparator and approximate information searcher 133 with the acoustic signal being currently input that has been fed back. Correction is made based on the difference from the characteristics to obtain the final all-band common quantization accuracy value.

次に、符号化情報量集計部１５及び周波数帯域決定部１６について、更に詳細に説明する。図２に示した周波数帯域決定部１６は、従来の符号化情報量を制御する量子化の繰り返し処理の役目を果たす部分である。現ブロックの音響信号において適当であると推定された全帯域共通量子化精度値と個別帯域量子化精度値を用いて量子化された周波数信号は、従来と同様の周波数帯域を確保しなくては品質上劣化が検知されてしまう。従って、音響信号の重要な要素である低中域を重視し、低域から順次加算された符号化情報量が目標とする割り当て情報量を超えるまで符号化を行えば、品質的に満足する符号化情報が生成できるはずである。   Next, the encoded information amount totaling unit 15 and the frequency band determining unit 16 will be described in more detail. The frequency band determination unit 16 shown in FIG. 2 is a part that plays a role of a conventional iterative process of quantization for controlling the amount of encoded information. The frequency signal quantized using the all-band common quantization accuracy value and the individual-band quantization accuracy value estimated to be appropriate in the current block acoustic signal must ensure the same frequency band as before. Quality degradation will be detected. Therefore, if the emphasis is on the low and middle range, which is an important element of the acoustic signal, and encoding is performed until the encoded information amount sequentially added from the low range exceeds the target allocated information amount, the code satisfying quality It should be possible to generate information.

与えられた符号化情報量に応じて可能な限り高域の周波数帯まで符号化してもよいが、符号化レートによって音質上過剰品質となる場合も有り得るので、適当な高域の周波数帯で打ち切り、余剰情報量は余剰情報量を蓄積するバッファに預ける事によって、現ブロック以降のブロック用に確保しておけばよい。このようなブロック間の情報量の貸し借りによって全体の符号化品質も向上する。   Depending on the amount of encoded information, encoding may be performed up to as high a frequency band as possible, but there may be an excess quality in terms of sound quality depending on the encoding rate. The surplus information amount may be reserved for blocks after the current block by depositing the surplus information amount in a buffer for accumulating the surplus information amount. By lending and borrowing the amount of information between blocks, the overall coding quality is also improved.

但し、全帯域共通量子化精度値の推定に誤差が生じた場合、情報量が不足し所望の周波数帯より低い帯域で打ち切られる恐れがある。そこで、図２の割り当て情報決定部１４では二種類の割り当て情報量を用意し、符号化における品質劣化を防いでいる。   However, if an error occurs in the estimation of the quantization accuracy value common to all bands, the amount of information is insufficient, and there is a possibility that the band is cut off in a band lower than the desired frequency band. Therefore, the allocation information determination unit 14 of FIG. 2 prepares two types of allocation information amounts to prevent quality degradation in encoding.

図５は二種類の割り当て情報量を用い、ブロック間でほぼ一定の品質の維持を可能にする、符号化情報量集計部１５及び周波数帯域決定部１６の動作説明用フローチャートである。まず、量子化された信号の分割周波数帯０〜ｎまでのｎ＋１個の分割周波数帯の符号化情報量Ｓを集計する（ステップＳ２１）。上記の変数ｎの初期値は０であり、最初は０番目の分割周波数帯、すなわち、最も低域の分割周波数帯の符号化情報量Ｓが集計される。   FIG. 5 is a flowchart for explaining the operation of the coded information amount totaling unit 15 and the frequency band determining unit 16 that uses two types of allocation information amounts and enables maintaining a substantially constant quality between blocks. First, the encoded information amount S of n + 1 divided frequency bands from the divided frequency bands 0 to n of the quantized signal is tabulated (step S21). The initial value of the variable n is 0, and the encoded information amount S of the 0th divided frequency band, that is, the lowest divided frequency band is first tabulated.

続いて、集計された符号化情報量Ｓが二種類の割り当て情報量のうち、少ない方の第１の割り当て情報量よりも小であるかどうか判定される（ステップＳ２２）。最初はｎ＝０であり、集計された符号化情報量Ｓは第１の割り当て情報量よりも小であるので、０番目の分割周波数帯に基づく周波数帯域を仮決定し（ステップＳ２３）、周波数帯域は適当かどうか判定する（ステップＳ２４）。ここで、周波数帯域が適当かどうかは、情報量の不足により符号化における品質劣化が生じないかどうかにより判断される。最初は当然周波数帯域が狭すぎるので、適当でないと判定され、周波数帯域ｎを１つ増加させ（ステップＳ２５）、今度は０番目の分割周波数帯域とそれより周波数帯域の高い１番目の分割周波数帯域の総符号化情報量Ｓが算出される（ステップＳ２１）。   Subsequently, it is determined whether or not the total encoded information amount S is smaller than the smaller one of the two types of allocation information amounts (step S22). Since n = 0 initially and the total encoded information amount S is smaller than the first allocation information amount, a frequency band based on the 0th divided frequency band is provisionally determined (step S23). It is determined whether the band is appropriate (step S24). Here, whether or not the frequency band is appropriate is determined based on whether or not quality degradation in encoding occurs due to an insufficient amount of information. Since the frequency band is naturally too narrow at first, it is determined that the frequency band is not appropriate, and the frequency band n is increased by one (step S25). This time, the zeroth divided frequency band and the first divided frequency band having a higher frequency band than this are divided. Is calculated (step S21).

以下、同様にして、低域の分割周波数帯から順次高域の分割周波数帯へ周波数帯域を広げていき、それらの符号化情報量Ｓを算出していき、周波数帯域ｎまでの総情報量が割り当て情報量以内であり、かつ、符号化される周波数帯域が所定の周波数帯域を満足すると、つまり、ステップＳ２４で周波数帯域が適当であると判定されると、情報割り当てを打ち切り、最終的な周波数帯域を決定し（ステップＳ２９）、量子化部の処理を終了する。   Hereinafter, in the same manner, the frequency band is sequentially expanded from the low frequency division frequency band to the high frequency division frequency band, the encoded information amount S is calculated, and the total information amount up to the frequency band n is If it is within the allocated information amount and the frequency band to be encoded satisfies the predetermined frequency band, that is, if it is determined in step S24 that the frequency band is appropriate, the information allocation is terminated, and the final frequency A band is determined (step S29), and the processing of the quantization unit is terminated.

周波数帯域が狭く不適当である場合、周波数帯域ｎを増加させながら割り当て情報量と帯域ｎまでの総符号化情報量Ｓとを比較し、可能な限り適当とされる所定の周波数帯域まで拡大する。適当とされる周波数帯域まで符号化情報量を取り込めればよいが、そのブロックにおける音響信号の符号化難易度が高く、符号化するための情報量が多く必要なとき、又は全帯域共通量子化精度値の推定値に誤差が存在し、量子化信号の情報量が増大したときなどは、ステップＳ２２で符号化情報量Ｓが割り当て情報量以上と判定される。   When the frequency band is narrow and inappropriate, the allocated information amount is compared with the total encoded information amount S up to the band n while increasing the frequency band n, and the frequency band is expanded to a predetermined frequency band as appropriate as possible. . It is sufficient if the amount of encoded information can be captured up to an appropriate frequency band. However, when the degree of difficulty in encoding an acoustic signal in the block is high and a large amount of information is required for encoding, or all-band common quantization When there is an error in the estimated value of the accuracy value and the amount of information of the quantized signal increases, it is determined in step S22 that the encoded information amount S is greater than or equal to the allocated information amount.

この場合は、補助情報量があるかどうか判定し（ステップＳ２６）、補助情報量が無ければ、そこで周波数帯域の決定処理を強制的に終了する（ステップＳ２７）。他方、補助情報量がある場合は、今までの第１の割り当て情報量に、周波数帯域を拡大するために余剰情報量として確保した補助情報量を追加して第２の割り当て情報量とし（ステップＳ２８）、この第２の割り当て情報量と符号化情報量Ｓとを再度比較する（ステップＳ２１）。これにより、符号化情報量Ｓが第２の割り当て情報量よりも小さくなると、周波数帯域を仮決定し、周波数帯域が適当であれば、周波数帯域を最終決定する（ステップＳ２３、Ｓ２４、Ｓ２９）。   In this case, it is determined whether or not there is an auxiliary information amount (step S26). If there is no auxiliary information amount, the frequency band determination process is forcibly terminated (step S27). On the other hand, if there is an auxiliary information amount, the auxiliary information amount secured as the surplus information amount for expanding the frequency band is added to the first allocation information amount so far to obtain the second allocation information amount (step In S28, the second allocation information amount and the encoded information amount S are compared again (step S21). Thereby, when the encoded information amount S becomes smaller than the second allocation information amount, the frequency band is provisionally determined. If the frequency band is appropriate, the frequency band is finally determined (steps S23, S24, S29).

このようにして、本実施の形態では、あるブロックにおける音響信号の符号化難易度が高く、符号化するための情報量が多く必要なとき、又は全帯域共通量子化精度値の推定値に誤差が存在し、量子化信号の情報量が増大したときなどにおいて、周波数帯域が他のブロックと比較し極端に低下する事態を防止することができる。   In this way, in this embodiment, when the degree of difficulty in encoding an acoustic signal in a certain block is high and a large amount of information is required for encoding, or there is an error in the estimated value of the all-band common quantization accuracy value. Thus, when the amount of information of the quantized signal increases, it is possible to prevent a situation where the frequency band is extremely lowered compared to other blocks.

また、本実施の形態では、割り当て情報量は１ブロックで消費してもよい通常の第１の割り当て情報量と、上記のような品質に関わる事態に対処するために余剰情報量として確保した補助情報量を追加した第２の割り当て情報量とを設定し、符号化情報量と品質保持の両立を図る役目を担っている。   In the present embodiment, the allocation information amount is a normal first allocation information amount that may be consumed in one block, and an auxiliary information amount that is secured as a surplus information amount in order to cope with the above-described quality-related situation. The second allocation information amount to which the information amount is added is set to play a role of achieving both the encoded information amount and the quality maintenance.

また、ここで得られた割り当て情報量と周波数帯域情報は、次ブロックの全帯域共通量子化精度値の推定材料として、図４の符号化情報蓄積器１３５に供給して使用する事もできる。このような再帰処理によって安定した品質が保たれ、更に信頼性の高い量子化精度で符号化を行えるため、情報量の余分な消費も抑制できる。   Also, the allocation information amount and frequency band information obtained here can be supplied to the encoded information storage unit 135 of FIG. 4 and used as an estimation material of the all-band common quantization accuracy value of the next block. Stable quality is maintained by such recursive processing, and encoding can be performed with more reliable quantization accuracy, so that excessive consumption of information can be suppressed.

なお、上記した符号化装置の機能をプログラムによりコンピュータに実現させるようにしてもよい。このプログラムは、記録媒体から読みとられてコンピュータに取り込まれてもよいし、通信ネットワークを介して伝送されてコンピュータに取り込まれてもよい。   Note that the functions of the above-described encoding device may be realized by a computer by a program. This program may be read from a recording medium and loaded into a computer, or may be transmitted via a communication network and loaded into a computer.

一般的な音響信号符号化装置のブロック図である。It is a block diagram of a general acoustic signal encoding device. 本発明の音響信号符号化装置の一実施の形態における量子化部のブロック図である。It is a block diagram of the quantization part in one Embodiment of the acoustic signal encoding apparatus of this invention. 図２の量子化部の動作説明用フローチャートである。It is a flowchart for operation | movement description of the quantization part of FIG. 図２中の全帯域共通量子化精度決定部の一実施の形態のブロック図である。FIG. 3 is a block diagram of an embodiment of an all-band common quantization accuracy determination unit in FIG. 2. 本発明における量子化部内の要部の動作説明用フローチャートである。It is a flowchart for operation | movement description of the principal part in the quantization part in this invention. 一重の繰り返し構造を持つ従来の音響信号符号化装置の量子化部の動作説明用フローチャートである。It is a flowchart for operation | movement description of the quantization part of the conventional acoustic signal encoding apparatus which has a single repetition structure.

符号の説明Explanation of symbols

１ 時間−周波数変換部
２ 周波数帯重み付け情報算出部
３、１０ 量子化部
４ 符号化信号生成部
１１ 量子化器
１２ 個別帯域量子化精度決定部
１３ 全帯域共通量子化精度決定部
１４ 割り当て情報決定部
１５ 符号化情報量集計部
１６ 周波数帯域決定部
１３１ 音響信号特性抽出部
１３２ 統計的音響信号特性分類情報
１３３ 特性比較器及び近似情報検索器
１３４ 全帯域共通量子化精度特定器
１３５ 符号化情報蓄積器


DESCRIPTION OF SYMBOLS 1 Time-frequency conversion part 2 Frequency band weighting information calculation part 3, 10 Quantization part 4 Coded signal generation part
DESCRIPTION OF SYMBOLS 11 Quantizer 12 Individual band quantization accuracy determination part 13 All-band common quantization accuracy determination part 14 Allocation information determination part 15 Encoded information amount totaling part 16 Frequency band determination part 131 Acoustic signal characteristic extraction part 132 Statistical acoustic signal characteristic Classification information 133 Characteristic comparator and approximate information searcher 134 All-band common quantization accuracy specifying unit 135 Encoded information storage unit


Claims (5)

時間領域のディジタル化された音響信号をある定められた時間間隔でブロック化した後、ブロック内の前記音響信号を複数の周波数帯域に分割し、分割した各周波数帯域毎の重み付けに基づき、ブロック内の前記音響信号の量子化を行い情報圧縮する音響信号符号化装置において、
前記複数の周波数帯域からなる全周波数帯域共通の全帯域共通量子化精度値を決定する全帯域共通量子化精度決定部と、
分割した各周波数帯域毎の聴覚的な重み付けによって、前記全帯域共通量子化精度値を各周波数帯域毎に修正するための個別帯域量子化精度値を決定する個別帯域量子化精度決定部と、
前記全帯域共通量子化精度値及び個別帯域量子化精度値に従い、前記ブロック内の音響信号を量子化する量子化器と、
情報圧縮率及び前記ディジタル化された音響信号の特性に従い、前記ブロック毎に情報量を割り当てる割り当て情報量決定部と、
前記量子化器から出力された量子化信号の符号化情報量を集計し、集計した符号化情報量に基づいて前記量子化信号の周波数帯域を決定する符号化情報量集計・周波数帯域決定部とを備え、
前記符号化情報量集計・周波数帯域決定部は、前記量子化器から出力された量子化信号における前記分割した各周波数帯域のうち、最も低域の分割周波数帯域の符号化情報量を集計し、前記情報量決定部によって割り当てられた情報量と前記最も低域の分割周波数帯域の符号化情報量とを比較し、前記最も低域の分割周波数帯域の符号化情報量が前記割り当てられた情報量より小さければ前記最も低域の分割周波数帯域に基づく周波数帯域を仮決定し、仮決定した周波数帯域が適当であるか否かをその周波数帯域の情報量が不足であるか否かにより判断し、適当でなければ次に高域の分割周波数帯域へ周波数帯域を広げて、その広げた周波数帯域における符号化情報量を集計し、その広げた周波数帯域における符号化情報量が前記割り当てられた情報量以内で、かつ、前記広げた周波数帯域が適当であると判断されるまで、前記最も低域の分割周波数帯域から高域の分割周波数帯域まで順次に周波数帯域を広げて、周波数帯域を最終決定することを特徴とする音響信号符号化装置。 After the time-domain digitized acoustic signal is blocked at a predetermined time interval, the acoustic signal in the block is divided into a plurality of frequency bands, and the blocks are divided into weights based on the divided weights for each frequency band. In the acoustic signal encoding apparatus that quantizes the acoustic signal and compresses the information,
An all-band common quantization accuracy determining unit that determines an all-band common quantization accuracy value that is common to all frequency bands including the plurality of frequency bands;
An individual band quantization accuracy determination unit that determines an individual band quantization accuracy value for correcting the all-band common quantization accuracy value for each frequency band by auditory weighting for each divided frequency band;
A quantizer for quantizing the acoustic signal in the block according to the all-band common quantization accuracy value and the individual band quantization accuracy value ;
According characteristic information compression ratio and the digitized audio signal, and allocation information amount determining unit to assign the amount of information for each said block,
An encoded information amount totaling / frequency band determining unit that aggregates the encoded information amount of the quantized signal output from the quantizer and determines the frequency band of the quantized signal based on the aggregated encoded information amount; With
The encoded information amount totaling / frequency band determining unit totals the encoded information amount of the lowest divided frequency band among the divided frequency bands in the quantized signal output from the quantizer, The information amount allocated by the information amount determination unit is compared with the encoded information amount of the lowest divided frequency band, and the encoded information amount of the lowest divided frequency band is the allocated information amount If smaller, tentatively determine a frequency band based on the lowest divided frequency band, determine whether the tentatively determined frequency band is appropriate or not by determining whether the amount of information in the frequency band is insufficient, If not appropriate, the frequency band is expanded to the next higher frequency band, and the amount of encoded information in the expanded frequency band is tabulated, and the amount of encoded information in the expanded frequency band is assigned to the allocated frequency band. The frequency band is sequentially expanded from the lowest divided frequency band to the higher divided frequency band until it is determined that the widened frequency band is appropriate within the specified amount of information. An acoustic signal encoding apparatus characterized by final determination . 前記符号化情報量集計・周波数帯域決定部は、前記最も低域の周波数帯域及び前記広げた周波数帯域における符号化情報量が前記割り当てられた情報量以内でない場合に、他のブロックで余剰した余剰情報量を補助情報として前記割り当てられた情報量に追加することを特徴とする請求項１記載の音響信号符号化装置。The coded information amount totaling / frequency band determining unit is configured so that, when the coded information amount in the lowest frequency band and the expanded frequency band is not within the allocated information amount, surplus in other blocks The acoustic signal encoding apparatus according to claim 1, wherein an information amount is added to the allocated information amount as auxiliary information. 前記全帯域共通量子化精度決定部は、前記ブロック内の音響信号エネルギー、音響信号振幅値及び割り当て情報量のうちの少なくとも一の情報を要素として予め求めた全帯域共通量子化精度値の統計情報に基づく前記全帯域共通量子化精度値と、現在符号化中の現ブロックの音響信号よりも過去のブロックの音響信号の符号化の際に用いた前記全帯域共通量子化精度値とのいずれか一方又は両者の組み合わせ値を、前記全帯域共通量子化精度値として決定することを特徴とする請求項１又は２記載の音響信号符号化装置。 The all-band common quantization accuracy determination unit is a statistical information of the all-band common quantization accuracy value obtained in advance using at least one of the acoustic signal energy, the acoustic signal amplitude value, and the allocated information amount in the block as an element. Any one of the all-band common quantization accuracy value based on the current block and the all-band common quantization accuracy value used when encoding the acoustic signal of the past block rather than the acoustic signal of the current block being encoded one or a combination value of both the acoustic signal encoding apparatus according to claim 1 or 2, wherein the determining the full band common quantization precision value. 前記全帯域共通量子化精度決定部は、
現ブロックの音響信号のブロック内の音響信号エネルギー、最大または代表的な振幅値、及びブロック内割り当て情報量の何れか一つ又は二つ以上を組み合わせた情報を、前記現ブロックの音響信号から抽出する音響信号特性抽出部と、
前記ブロック内の音響信号のエネルギー、最大または代表的な音響信号の振幅値、及びブロック割り当て情報量の何れか一つ又は二つ以上を組み合わせて得られた全帯域共通量子化精度値の統計結果を示す統計的音響信号特性分類情報と、前記音響信号特性抽出部から出力された情報とを比較し、前記統計的音響信号特性分類情報の中から前記音響信号特性抽出部から出力された情報に最も近似する統計的音響信号特性分類情報を検索する特性比較部及び近似情報検索部と、
前記検索結果から現ブロックの全帯域共通量子化精度値を特定する全帯域共通量子化精度特定部と
を有することを特徴とする請求項１又は２記載の音響信号符号化装置。 The all-band common quantization accuracy determination unit
Extracts information from the acoustic signal of the current block from the acoustic signal of the current block, by combining one or more of the acoustic signal energy, the maximum or representative amplitude value in the block, and the amount of allocated information in the block. An acoustic signal characteristic extraction unit for
Statistical results of all-band common quantization accuracy values obtained by combining any one or more of the energy of the acoustic signal in the block, the amplitude value of the maximum or representative acoustic signal, and the block allocation information amount The statistical acoustic signal characteristic classification information indicating the information output from the acoustic signal characteristic extraction unit is compared with the information output from the acoustic signal characteristic extraction unit from the statistical acoustic signal characteristic classification information A characteristic comparison unit and an approximate information retrieval unit for retrieving the most approximate statistical acoustic signal characteristic classification information;
The search result from the acoustic signal encoding apparatus according to claim 1, wherein further comprising a full band common quantization accuracy specifying unit for specifying a full band common quantization precision values of the current block. 前記全帯域共通量子化精度決定部は、過去に決定した前記全帯域共通量子化精度値を用いて量子化された量子化信号の符号化情報量と、過去の１又は複数ブロックの前記量子化信号の周波数帯域の情報とを再帰的に用いて、前記全帯域共通量子化精度特定部で特定された前記全帯域共通量子化精度値を補正する補正部を含むことを特徴とする請求項４記載の音響信号符号化装置。 The all-band common quantization accuracy determination unit includes an encoded information amount of a quantized signal quantized using the all-band common quantization accuracy value determined in the past, and the past one or more blocks of the quantization and information signal in the frequency band used recursively, claim 4, characterized in that it comprises a correcting unit for correcting the said full band common quantization precision values identified in the entire band common quantization accuracy specification unit The acoustic signal encoding device described.
JP2004184829A 2004-06-23 2004-06-23 Acoustic signal encoding device Active JP4301091B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2004184829A JP4301091B2 (en) 2004-06-23 2004-06-23 Acoustic signal encoding device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP2004184829A JP4301091B2 (en) 2004-06-23 2004-06-23 Acoustic signal encoding device

Publications (2)

Publication Number Publication Date
JP2006010817A JP2006010817A (en) 2006-01-12
JP4301091B2 true JP4301091B2 (en) 2009-07-22

Family

ID=35778182

Family Applications (1)

Application Number Title Priority Date Filing Date
JP2004184829A Active JP4301091B2 (en) 2004-06-23 2004-06-23 Acoustic signal encoding device

Country Status (1)

Country Link
JP (1) JP4301091B2 (en)

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4548348B2 (en) * 2006-01-18 2010-09-22 カシオ計算機株式会社 Speech coding apparatus and speech coding method
JP2009086661A (en) * 2007-09-12 2009-04-23 Kawai Musical Instr Mfg Co Ltd Information compression method of musical sound waveform, information decompression method, computer program for information compression, information compression device, information decompression device, and data structure
JP4924746B2 (en) * 2010-08-13 2012-04-25 カシオ計算機株式会社 Code amount control device, code amount control method, code amount control program, moving image recording device
JP5800920B2 (en) * 2012-02-07 2015-10-28 日本電信電話株式会社 Encoding method, encoding apparatus, decoding method, decoding apparatus, program, and recording medium
JP5786044B2 (en) * 2012-02-07 2015-09-30 日本電信電話株式会社 Encoding method, encoding apparatus, decoding method, decoding apparatus, program, and recording medium
WO2013187498A1 (en) * 2012-06-15 2013-12-19 日本電信電話株式会社 Encoding method, encoding device, decoding method, decoding device, program and recording medium

Also Published As

Publication number Publication date
JP2006010817A (en) 2006-01-12

Similar Documents

Publication Publication Date Title
US8244524B2 (en) SBR encoder with spectrum power correction
CN1702974B (en) Method and apparatus for encoding/decoding a digital signal
CA2352416C (en) Audio encoder and psychoacoustic analyzing method therefor
JP5704397B2 (en) Encoding apparatus and method, and program
US20070016404A1 (en) Method and apparatus to extract important spectral component from audio signal and low bit-rate audio signal coding and/or decoding method and apparatus using the same
JPH05304479A (en) High efficient encoder of audio signal
US6240388B1 (en) Audio data decoding device and audio data coding/decoding system
KR101361933B1 (en) Frequency band scale factor determination in audio encoding based upon frequency band signal energy
US20060053006A1 (en) Audio encoding method and apparatus capable of fast bit rate control
JP4301091B2 (en) Acoustic signal encoding device
JP3519859B2 (en) Encoder and decoder
US8576910B2 (en) Parameter selection method, parameter selection apparatus, program, and recording medium
EP1139336A2 (en) Determination of quantizaion coefficients for a subband audio encoder
KR20160120713A (en) Decoding device, encoding device, decoding method, encoding method, terminal device, and base station device
JP2006145782A (en) Encoding device and method for audio signal
JP2006018023A (en) Audio signal coding device, and coding program
US20130346073A1 (en) Audio encoder/decoder apparatus
KR20170089982A (en) Signal encoding and decoding method and devices
US6678653B1 (en) Apparatus and method for coding audio data at high speed using precision information
JP2000151413A (en) Method for allocating adaptive dynamic variable bit in audio encoding
JP4062971B2 (en) Audio signal encoding method
JP4822816B2 (en) Audio signal encoding apparatus and method
JP4301092B2 (en) Acoustic signal encoding device
JP2001242895A (en) Audio encode device and audio encode method
JP2005004119A (en) Sound signal encoding device and sound signal decoding device

Legal Events

Date Code Title Description
A621 Written request for application examination

Free format text: JAPANESE INTERMEDIATE CODE: A621

Effective date: 20060630

A977 Report on retrieval

Free format text: JAPANESE INTERMEDIATE CODE: A971007

Effective date: 20081111

A131 Notification of reasons for refusal

Free format text: JAPANESE INTERMEDIATE CODE: A131

Effective date: 20081202

A521 Request for written amendment filed

Free format text: JAPANESE INTERMEDIATE CODE: A523

Effective date: 20090129

TRDD Decision of grant or rejection written
A01 Written decision to grant a patent or to grant a registration (utility model)

Free format text: JAPANESE INTERMEDIATE CODE: A01

Effective date: 20090331

A01 Written decision to grant a patent or to grant a registration (utility model)

Free format text: JAPANESE INTERMEDIATE CODE: A01

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20120501

Year of fee payment: 3

R151 Written notification of patent or utility model registration

Ref document number: 4301091

Country of ref document: JP

Free format text: JAPANESE INTERMEDIATE CODE: R151

A61 First payment of annual fees (during grant procedure)

Free format text: JAPANESE INTERMEDIATE CODE: A61

Effective date: 20090413

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20120501

Year of fee payment: 3

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20120501

Year of fee payment: 3

S111 Request for change of ownership or part of ownership

Free format text: JAPANESE INTERMEDIATE CODE: R313111

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20120501

Year of fee payment: 3

R350 Written notification of registration of transfer

Free format text: JAPANESE INTERMEDIATE CODE: R350

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20120501

Year of fee payment: 3

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20130501

Year of fee payment: 4